🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the FuriosaAI RNGD run OpenELM-3B?
Yes — here's how fast and up to what context, with real measured numbers.
OpenELM-3B on the FuriosaAI RNGD
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 558k | 160.4-194.9 t/s | ~4.3-5.2 (slow) |
| Q8 | yes | 605k | 320.7-389.8 t/s | ~8.6-10.4 (slow) |
| Q4 | yes | 625k | 572.7-696.1 t/s | ~15.3-18.6 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to system RAM offload (depends on your system RAM) and speed drops to the "beyond" figure.
FAQ
Can the FuriosaAI RNGD run OpenELM-3B?▶
Yes. At Q4 it generates ~572.7-696.1 tok/s and fits up to 625k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~15.3-18.6 tok/s.
How much context fits?▶
At Q4, up to about 625k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.